| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Obstetrics and Gynecology (C.W., J.E.P., S.V.P., J.K., L.L., E.G.L., J.W.L.), Center for Reproductive Sciences (L.L., J.W.L.), and Department of Medicine (R.K.T.), Columbia University, New York, New York 10032; Serono Reproductive Biology Institute (S.P.), Rockland, Massachusetts 02370; and Department of Biostatistics (R.T.O.), School of Public Health, Columbia University, New York, New York 10032
Address all correspondence and requests for reprints to: Dr. J. W. Lustbader, Columbia University Medical Center, 630 West 168th Street, Department of Obstetrics and Gynecology, P & S 16-408, New York, New York 10032. E-mail: jwl2{at}columbia.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
FSH, LH, TSH, and placental-derived human chorionic gonadotropin (hCG) are members of the glycoprotein hormone family. All family members are composed of two noncovalently bound subunits. The 
-subunit is shared by all family members, whereas the ß-subunit is unique to each hormone and confers biological specificity (7). hCG has the longest circulating half-life of the members due to the presence of a unique carboxy-terminal peptide (CTP) on the ß-subunit (8). Glycosylation of the CTP occurs at four O-linked carbohydrate attachment sites (9) and increases the terminal (elimination) half-life (t1/2) of hCG by reducing glomerular filtration within the kidney (10, 11, 12, 13). The CTP has also been shown to increase the half-life of recombinant human FSH (rhFSH) when covalently bound to the ß-subunit (rhFSH-CTP) (14, 15).
rhFSH-CTP has recently been evaluated for human applications and was found to have a t1/2 two to three times longer than native rhFSH in hypogonadotropic hypogonadal men after sc injection (16). Furthermore, in women, a single administration of rhFSH-CTP was sufficient to induce the growth of multiple follicles with a parallel rise in serum inhibin, a marker of follicular growth produced by granulosa cells (17, 18). The first live birth of a healthy infant was recently reported using an ovarian stimulation protocol incorporating rhFSH-CTP in combination with native rhFSH in a protocol requiring fewer injections than standard rhFSH-only regimens (19).
Efforts to further increase the in vivo bioactivity of FSH through the addition of two tandem CTP signal sequences showed no effect (14). The signal sequence for O-linked glycosylation is poorly defined, whereas N-linked glycosylation sites are specifically defined by the trimer peptide sequence Asn-X-Thr/Ser, where X can be any amino acid residue except proline. N-linked glycosylation sites can be introduced in tandem into FSH to generate a series of FSH analogs with variable, increased half-life and potentially enhanced in vivo bioactivity.
Previously, we reported the characterization and biological activity of a single- chain rhFSH construct, rhFSH-N2, that placed a synthetic polypeptide encoding two additional N-linked glycosylation sites between the ß- and -FSH subunits (20). The addition of the N-linked carbohydrate moieties did not interfere with proper folding of the hormone or binding to its receptor. rhFSH-N2 had similar in vitro activity to rhFSH and rhFSH-CTP and, like rhFSH-CTP, a t1/2 that was 2-fold longer than rhFSH. Furthermore, other researchers have described increasing the half-life of rhFSH by introducing two N-linked glycosylation sites at the N terminus of the -subunit (21). These results suggest that added N-glycosylation sites can increase half-life, regardless of the locations of the added sites within the hormone analog. Thus, novel FSH analogs containing variable numbers of N- (and O-) linked carbohydrates were examined in this report to determine how altering the number and type of additional carbohydrate moieties affects pharmacokinetic and pharmacodynamic properties of FSH. We found that increasing the number of glycosylation sites, either O-linked or N-linked, increased the t1/2, but there was a maximum t1/2 such that further glycosylation provided no additional benefit to prolongation of the half-life. However, when we compared the activities of these FSH analogs in vivo using several criteria that included comparisons of ovarian weights, inhibin A production, and follicle development, we found that N-linked analogs were significantly more biopotent than O-linked analogs.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We previously described rhFSH-N2, an expression vector that encodes the ß- and -subunits of FSH conjoined by a sequence encoding two N-linked carbohydrate attachment sites (20). All constructs described in this report were derived from rhFSH-N2. The N2 linker sequence between FSHß and FSH subunits was excised using the unique BamH1 and Spe1 restriction sites present in the construct. New carbohydrate linker sequences, created through additional enzyme digestions or site-directed mutagenesis followed by PCR, were reintroduced into the original FSH plasmid by ligation. Fidelity of the new vector was confirmed by sequencing. Figure 1
shows the linker sequences for all other analogs described in this report.
|
FSH constructs and pSV2neo, an expression vector containing the selectable marker neomycin phosphotransferase, were cotransfected into CHO-K1 cells in the presence of Lipofectamine Plus Reagent (Invitrogen, Carlsbad, CA). G418-resistant colonies were selected in Ham’s F12 medium containing 10% fetal calf serum (Invitrogen), 100 U/ml penicillin, 100 µg/ml streptomycin, 4 mM glutamine, and 500 µg/ml G418 (Gemini Bioproducts, Woodland, CA) for 6–7 d and subsequently pooled together and maintained in the same medium. FSH production by pooled clones was detected by RIA using an antibody to ß-FSH (Biodesign International, Saco, ME) as previously described (20). FSH-containing supernatants were obtained from cellular suspensions in spinner bottles that were initially seeded at 1 x 105 cells/ml into CHO-S-SFM II medium (Invitrogen) containing 400 µg/ml G418. Cells usually reached a density of 2 x 106 cells/ml on d 6 or 7, at which time cell supernatant was harvested on d 7 or 8. Supernatants were supplemented with 0.2 mM phenylmethylsulfonylfluoride, filtered through a 0.2-µm membrane, and kept at 4 C until purification.
Protein purification of rhFSH analogs from CHO-K1 supernatant
rhFSH analogs were purified by adsorption and elution from an -chain-specific monoclonal antibody column that was prepared by coupling 5 mg purified A103 Ig/ml to CNBr-Sepharose-4B (Amersham Biosciences, Piscataway, NJ) according to the manufacturer’s instructions (22). After applying the cell supernatant, the column was washed in 50 bed volumes of PBS followed by 2 bed volumes of distilled water. The rhFSH analog was eluted in 3–4 bed volumes of 1 M acetic acid and immediately dried en vacuo in a Speed-Vac (Savant Instruments, Holbrook, NY).
Matrix-assisted laser desorption ionization (MALDI) analysis
rhFSH analogs, prepared specifically for MALDI analyses, were purified by reverse-phase HPLC using a Hewlett-Packard 1090 (Hewlett-Packard, Palo Alto, CA) and a Vydac C4 column (Grace Vydac, Hesperia, CA). Twenty-five percent of peak fractions were dried en vacuo and dissolved in 3 µl of a saturated solution of -cyano-4-hydroxycinnamic acid in formic acid-propanol-water (1:2:3). A 0.5-µl aliquot was spotted on a sample plate prepared with a thin layer of -cyano-4-hydroxycinnamic acid for mass spectrometric analyses in a PerSeptive Biosystems Voyager-DE RP MALDI mass spectrometer (PerSeptive Biosystems, Inc., Framingham MA).
Electrophoresis and Western blotting
rhFSH (Gonal-F; Serono, Inc. Rockland, MA), single-chain rhFSH-CTP, rhFSH-N0, rhFSH-N1, rhFSH-N2, rhFSH-N4, and rhFSH-O1 were reduced and electrophoresed through an SDS-PAGE gel and silver stained (23). Another set of nonreduced samples were electrophoresed and transferred to nitrocellulose using standard techniques (23, 24, 25). The nitrocellulose was incubated overnight in a 1:10,000 dilution of mouse antihuman FSHß monoclonal antibody (Biodesign International) after being blocked in 5% BSA for 1 h. The membrane was subsequently washed five times and incubated in 1:10,000 dilution of peroxidase-conjugated polyclonal antibody to mouse Ig (Amersham-Biosciences), for 1 h. Bands were visualized, after additional washes in Tris-buffered saline, by incubation in chemiluminescent detection reagent (Amersham-Biosciences) and exposed to x-ray film.
Isoelectric focusing gel electrophoresis
Isoelectric points (pI) of FSH proteins (400 pmol/lane) were determined by electrophoresis through Novex pre-cast IEF gels containing a pI range of pH3–pH7 (Invitrogen, Carlsbad, CA). Samples were visualized by silver stain.
In vitro biology
FSH concentration of all analogs was determined by RIA using a monoclonal antibody to FSHß (Biodesign International). FSH receptor binding and activation in vitro was assessed by measuring cAMP activity in CHO cells expressing the hFSH receptor (CHO-FSHR) as previously described (26). CHO-FSHR cells were detached from culture dishes using 0.02% versene. Twenty thousand cells per tube were incubated with varying amounts of standard FSH or FSH analogs for 15 min at 37 C. Tubes were incubated for 3 min at 75 C and then centrifuged for 10 min at 750 x g. One hundred microliters of supernatant were assayed for cAMP concentration using a commercial RIA kit (Perkin-Elmer Life Sciences, Boston, MA).
In vivo bioactivity and half-life determinations
Approval for animal experiments was obtained from the Institutional Animal Care and Use Committees at Serono (Rockland, MA) and Columbia University (New York, NY). Twenty-one-day-old female Sprague Dawley rats, obtained from Charles River Laboratories (Wilmington, MA), were housed three to five rats per cage and given standard food and tap water ad libitum. Animals were randomly assigned to receive a single injection of one of the following treatments either iv or ip: saline, rhFSH, rhFSH-N0, rhFSH-N1, rhFSH-N2, rhFSH-N4, rhFSH-O1, or rhFSH-CTP. There was no statistical difference in animal weights between each group. Each analog was injected at a dose of 42 IU (27, 28, 29), as determined by the FSH RIA.
Animals were anesthetized by inhalation of isofluorane before injection. Blood collection was obtained by periorbital venipuncture or tail bleeds. Rats were euthanized at the conclusion of the experiments by carbon dioxide inhalation followed by exsanguinations via cardiac puncture. FSH was detected in serum samples using RIA or Immulite (Diagnostic Products Corporation, Los Angeles, CA), and inhibin A levels were measured using an ELISA kit (Diagnostic Systems Laboratories, Webster, TX).
Pharmacokinetic analysis was performed using Winnonlin 1.0 software (Pharsight Corporation, Mountain View, CA) for iv injection and PK Solutions 2.0 software (Summit Research Services, Montrose, CO) for ip injections. For iv injections, mean values for each time point within a group were used to estimate pharmacokinetic parameters for t1/2 and area under the serum concentration time curve. For ip injections, pharmacokinetic parameters were determined for each individual rat.
Histological analysis
Rats were euthanized at 74 h post injection (ip), and ovaries were removed, weighed, fixed in paraformaldehyde, and embedded in paraffin. Each ovary was serial-sectioned completely at intervals of 10 µm. Mounted sections were stained in hematoxylin and eosin according to standard procedures. Ovaries were evaluated using a Leitz Dialux 20 light microscope equipped with a SPOT digital camera (Diagnostic Instrument, Inc., Sterling Heights, MI). Each section was evaluated for the presence of antral follicles. Follicles with clear antrum development, at stages 6–8 of the Pedersen and Peters grading system (30), were included in the follicle count. The ovary having the largest number of antral follicles within each treatment group was further evaluated for follicle size. Sizes of the first 30 (saline) or 50 (all other treatment groups) follicles with a clear antrum and the nucleus present in the section were measured using a Reichert eyepiece reticle, and these measurements were converted to actual follicle sizes per the manufacturer’s instructions (Reichert Scientific Instruments, Buffalo, NY).
Statistical analysis
Ovarian weights and inhibin levels were compared between groups in a one-way ANOVA model. The data were log transformed first to stabilize the variance between groups and satisfy the model assumptions because groups with higher means tended to have higher variances. The primary comparison of interest was between rhFSH-X (X = additional carbohydrate or tether) and rhFSH, and this difference was tested using a single contrast within the one-way structure. In addition, individual pharmacokinetic parameters (ip) were compared using a one-way ANOVA model. The comparisons of interest were each analog to FSH and additionally rhFSH-N4 vs. rhFSH-CTP and rhFSH-N4 vs. rhFSH-N2. These hypotheses were tested with simultaneous contrasts using a Bonferroni adjustment to control for multiple comparisons. Statistical significance of differences was assessed by ANOVA for one-way analysis or simultaneous comparison of FSH-N4 to all other groups using Dunnett’s method (31). P < 0.05 was considered significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
rhFSH-CTP, rhFSH-N0, rhFSH-O1, rhFSH-N1, rhFSH-N2, and rhFSH-N4 were produced by CHO cells at concentrations of 1.3–14 pmol/ml in suspension culture. Western blot analysis showed that rhFSH migrated at an apparent mass of 41,000 Da (Fig. 2B). No additional carbohydrate was added to rhFSH-N0. The increased molecular mass (46,000 Da) compared with rhFSH was attributed to the additional 20 amino acids in the linker sequence (Fig. 2 and Table 1). The hyperglycosylated FSH analogs migrated more slowly than rhFSH-N0 (rhFSH-N1, 52,000 Da; rhFSH-CTP, 54,000 Da; hFSH-N2, 56,000 Da; and rhFSH-N4, 58,000 Da). The apparent molecular mass of rhFSH-O1 (44,000 Da), containing only a modest increase in carbohydrate, was similar to rhFSH-N0. Glycosylation can increase the sialic acid content of a protein, resulting in a more acidic profile that is detectable by a decrease in pI. Figure 3 shows that dimeric rhFSH exhibits a pI between 4.75 and 3.9 (lane 4), whereas the hyperglycosylated isoforms all exhibited a lower pI range of 4.45–3.5, supporting the assertion that rhFSH-CTP, rhFSH-N2, rhFSH-N1, and rhFSH-N4 are more glycosylated than rhFSH. The pI ranges for rhFSH-O1 (lane 8), rhFSH-N0 (lane 1), and rhFSH (lane 4) were similar probably because these proteins either lack or contain limited additional carbohydrate. The same amount of protein was applied to each lane, yet some analogs focused less discreetly (lanes 2, 7, and 8), which may be attributable to a greater number of isoforms for those analogs.
|
|
|
). Noncovalently associated rhFSH - and ß-chains, for example, exhibited two peaks that were measured at 12,034 and 16,569 Da, respectively, for a combined dimer mass of 28,603 Da. The peaks for single-chain rhFSH-CTP and rhFSH-N2 were larger than the rhFSH heterodimer, as expected (36,971 and 38,514 Da, respectively). rhFSH-N4, having four potential N-linked glycosylation sites, exhibited the largest molecular mass at 40,925 Da, which was a 30% increase over the rhFSH. In vitro bioactivity
Increased glycosylation did not interfere with FSH activity because all FSH analogs stimulated the production of cAMP in CHO-K1 cells that constitutively expressed the FSH receptor (Fig. 4). rhFSH-N4, rhFSH-N1, rhFSH-O1, and rhFSH-N0 were able to stimulate a rise in cAMP levels comparable to rhFSH and other previously described analogs (rhFSH-N2 and rhFSH-CTP), indicating that the proteins were folded properly and were biologically active (Fig. 4).
|
Two sets of experiments were performed for half-life determination in vivo, wherein 22-d-old rats selected at random received single-dose injections of FSH analogs equivalent to 42 IU as assessed by RIA (27, 28, 29). In the first experiment, rats (n = 5 in each group) received a single dose of one of the following treatments through ip injection: saline, rhFSH, rhFSH-CTP, rhFSH-O1, rhFSH-N0, rhFSH-N1, rhFSH-N2, or rhFSH-N4. Blood was obtained by venipuncture via dorsal tail vein at the following intervals post injection: 3, 6, 9, 12, 24, 30, and 74 h. In the second experiment of a select group of analogs, rats (n = 3 in each group) received a single dose of one of the following treatments iv via dorsal tail vein: saline, rhFSH, rhFSH-CTP, rhFSH-N2, or rhFSH-N4. Blood was obtained by periorbital venipuncture at the following intervals post injection: 0.5, 1, 3, 6, 12, and 24 h. Rats were euthanized at 48 h post injection, and blood was collected by cardiac puncture. Serum FSH levels were measured as described in Materials and Methods.
Pharmacokinetic parameters for both iv and ip injection are depicted in Table 2. Additionally, mean concentration time curves for ip injection are illustrated in Fig. 5.
|
|
shows that the increase in glycosylation can enhance the t1/2 of the analogs. The t1/2 values for all analogs containing additional carbohydrate were significantly higher than rhFSH. However, the modest increase in t1/2 (ip) of rhFSH-O1 may be, in part, due to the linking of the subunits, given that the t1/2 (ip) of rhFSH-N0 was increased by a similar amount. The t1/2 (ip) of rhFSH-N1 was approximately 1.5 times higher than rhFSH. Although the t1/2 (iv/ip) of rhFSH-N4 was approximately 2-fold longer than the t1/2 of rhFSH (P < 0.001 for ip), it was similar to that of rhFSH-CTP and rhFSH-N2. Differences between t1/2 (ip) values for rhFSH-CTP, rhFSH-N2, and rhFSH-N4 were insignificant (P > 0.05). Thus, the addition of carbohydrate prolongs the half-life of FSH with either N-linked and O-linked analogs, but increasing the number of glycosylation sites beyond N2 appears not to be correlated with a longer t1/2. Pharmacodynamic studies
Potency of each analog was evaluated in vivo using several parameters that included ovarian weight augmentation, ovarian follicle-specific hormone levels (inhibin A), and ovarian follicle development. All FSH therapies were administered as a one-time injection at dose equivalence of 42 IU (see Materials and Methods). No other hormones were given during the experiment.
Ovarian weight augmentation
Figure 6 shows mean ovarian weights at 74 h post treatment. Groups stimulated by rhFSH-CTP, rhFSH-N2, or rhFSH-N4 had significantly higher ovarian weights compared with saline or rhFSH (P < 0.05). rhFSH-N0 and rhFSH-O1 were similar to rhFSH, whereas rhFSH-N1 showed a slightly increased weight. rhFSH-N4 stimulated a higher mean ovarian weight of 22.5 mg compared with all other treatments, but it was not statistically more effective than rhFSH-CTP (P > 0.05) or rhFSH-N2 (P > 0.05).
|
Secondary preantral follicles produce inhibin A in rats, and an increase of inhibin A levels correlates with an increase in follicular growth (32). We found that inhibin A levels for rhFSH- and saline-treated control animals were similar at 74 h (Fig. 7), presumably due to the short t1/2 of rhFSH at 6.3 h post ip injection (Table 2). To confirm that rhFSH was properly administered and biologically active, we measured inhibin A levels at 24 h and found that the inhibin A level was 60 times higher than at 74 h (data not shown). Hyperglycosylated forms of rhFSH induced inhibin A levels at 74 h that were significantly higher than the saline controls or rhFSH (Dunnet’s method, P < 0.05; Fig. 7). A comparison of the O-linked analogs showed that CTP, for example, was able to induce significantly higher inhibin levels in the rats compared with rhFSH-O1, which is consistent with the theory that additional glycosylation leads to increased biological activity. All N-linked analogs stimulated significantly higher inhibin levels compared with the O-linked rhFSH-CTP. Even rhFSH-N1 induced more than twice as high inhibin A levels than rhFSH-CTP, suggesting that FSH activity is more efficient when the hormone is modified by N-linked glycosylation. The mean inhibin A level in the rhFSH-N4-treated group was the highest among all treatment groups at 6369 pg/ml and was statistically equivalent to rhFSH-N2 (P = 0.3). The O-linked glycosylated analog, rhFSH-CTP, was much less efficient at inducing inhibin A (1638 pg/ml) compared with rhFSH-N4 (P < 0.001). FSH analogs that contained only one N-link or one O-link glycosylation site showed the same differential characteristic where inhibin A levels in rhFSH-N1-treated animals (at 4580 pg/ml) were 40 times higher than rhFSH-O1-treated animals (Fig. 7). We conclude from these data that N-linked glycosylated analogs stimulate a larger number of active follicles than their O-linked structural counterparts.
|
Antral follicle number and size was determined by examining hematoxylin and eosin-stained serial sections of ovaries from each treatment group (n = 4 ovaries in each group). Representative sections from each group at 74 h of exposure to analog are shown in Fig. 8. The saline-treated ovary shows a wide range of follicles from preantral to small, early antral follicles, whereas there were more antral follicles in all hormone-injected rats. Table 3 shows that the mean antral follicle count increased in all hormone-treated animals. rhFSH-N1 had 25% more antral follicles than rhFSH-CTP, but follicle size was similar for both analogs. rhFSH-N2- and rhFSH-N4-treated animals had twice as many antral follicles, and the mean size of the antral follicles were nearly 50% larger than rhFSH-treated animals (simultaneous P < 0.0001 using Dunnett’s method). rhFSH-N4-treated animals produced the highest number of large, preovulatory antral follicles, which was determined by scoring the percentage of follicles 
600 µm (saline, 0%; rhFSH, 0%; rhFSH-O1, 0%; rhFSH-CTP, 4%; rhFSH-N1, 2%; rhFSH-N2, 4%; and rhFSH-N4, 14%). The one-sided contrast showed that rhFSH-N4 did promote more follicle growth than rhFSH-N2.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We found that the addition of as many as four N-linked carbohydrate moieties to rhFSH did not interfere with protein folding or in vitro activation of the FSH receptor. Thus, it is possible to add multiple carbohydrate chains and mass to FSH without adversely affecting its biological activity. The upper limit of total carbohydrate mass that is permissive or conducive to hormone activity remains unknown.
Hyperglycosylation increased the half-life of FSH, but its effect was not linear with respect to carbohydrate size and number. rhFSH-N4 and two other (smaller) previously described hyperglycosylated proteins, rhFSH-CTP (14, 15) and rhFSH-N2 (20), had similar t1/2 kinetics. Thus, there does appear to be a maximum increase in half-life that can be achieved with glycosylation alone.
Previous pharmacology studies have shown that FSH is excreted by the kidney and that acidic endogenous isoforms of FSH are cleared more slowly than less acidic forms, regardless of carbohydrate type or amount (11, 12, 13). Our analysis shows that glycosylation lowers the pI profile and is most likely attributable to an increase in the sialic acid content of the carbohydrate side chains. Sialylation is known to confer a longer t1/2 upon native pituitary FSH isoforms in vivo (34), and our study showed that the half-lives of the more acidic analogs, rhFSH-CTP, rhFSH-N2, and rhFSH-N4, were twice as long as rhFSH. rhFSH-N0 and rhFSH-O1, in contrast, had pI profiles similar to rhFSH and half-lives only slightly longer than the native hormone. This increase may be attributable, in part, to the tether. The pI profile and the half-life of FSH-N1 were intermediate between rhFSH and the more heavily hyperglycosylated analogs. Furthermore, we have shown that rhFSH-N2 and rhFSH-N4, which are 5% and 9% larger than rhFSH-CTP, respectively, display similar half-lives (Table 1 and Fig. 2). Recently, it has been shown that additional glycosylation at the amino-terminal side of the -subunit also increases the t1/2, indicating that the position of the added carbohydrate does not influence the pharmacokinetics (21).
Two of these analogs, rhFSH-N2 and rhFSH-N4, induced more gonadotropic activity in vivo than rhFSH-CTP, which has four O-linked glycosylation sites. This effect may be due to several factors including monosaccharide composition and branching within the carbohydrate moieties, the nature of the glycosyl linkage, or because N-linked analogs have a higher carbohydrate mass than the O-linked rhFSH-CTP.
Prolonged t1/2 alone of our FSH analogs cannot adequately explain differences of in vivo bioactivity. The t1/2 of rhFSH-N1, for example, was shorter than rhFSH-CTP, but it induced 2.5 times more inhibin A (Fig. 7). Furthermore, rhFSH-N2 and rhFSH-N4 induced inhibin A levels three and four times higher than rhFSH-CTP, respectively, despite a negligible difference in t1/2. Thus, N-linked sugars augment gonadotropic activity independent of alterations in pharmacokinetic properties in a manner that is disproportionate to O-linked counterparts. The mechanism that is responsible for this effect is not known.
rhFSH-N2 and rhFSH-N4 stimulated the same number of antral follicles (Table 3), but rhFSH-N4 stimulated higher inhibin A levels (Fig. 7) and a higher percentage of follicles that grew to more than 600 µm (see Results). rhFSH-N2 and rhFSH-N4 have similar half-lives, similar pI profiles, and differ in size by only 6% (according to MALDI analysis and gel electrophoresis), leading us to hypothesize that rhFSH-N4 is a more efficient stimulator of follicle growth, possibly due to glycosyl chain length or steric hindrance. rhFSH-N2, for example, has only two additional glycosylation sites and may experience less spatial encumbrance than rhFSH-N4, thereby facilitating the addition of large, highly branched carbohydrates at each location. Alternatively, rhFSH-N4 has more carbohydrate moieties, and these chains may have fewer bifurcations due to increased spatial obstruction. This interpretation is supported by a study that showed an inverse relationship between FSH bioactivity and carbohydrate complexity (11).
Two recent clinical trials have shown that hyperglycosylation increases the half-life of FSH (19, 20), where the t1/2 of rhFSH-CTP was two to three times longer than rhFSH. An increase in inhibin B levels was detected in hypogonadotropic hypogonadal men injected with rhFSH-CTP, validating its effectiveness in this clinical population (16). We speculate that rhFSH-N4 will have enhanced bioactivity in men and women, enabling fewer injections with comparable or improved results. In particular, rhFSH-N4 might be adapted for stimulating optimal spermatogenesis in hypogonadotropic hypogonadal men who have typically required a prolonged therapy of follitropin injections in conjunction with hCG. These carbohydrate-rich FSH analogs may also have enhanced efficacy, as is illustrated by a case report of a woman who did not respond adequately to current rhFSH formulations but who was able to produce 10 large follicles after rhFSH-CTP administration, which culminated in the birth of a healthy infant (19).
Our studies have shown that FSH analogs containing multiple N-linked glycosylation sites, such as rhFSH-N2 and rhFSH-N4, are more biopotent than those with multiple O-glycosylation sites. The addition of multiple carbohydrate side chains does not appear to interfere with protein folding, and in fact, amplified glycosylation further enhances ovarian follicle growth. Furthermore, N-linked analogs can be designed more easily because signal sequences are known for N-linked glycosylation. Such FSH agonists represent a novel and potentially more powerful group of therapeutic candidates. Additional studies are needed to evaluate aspects of carbohydrate structure in bioactivity modulation. These include different branching patterns of carbohydrates in our analogs, determination of the optimal spacing for these glycosylation sequences, and the influence of additional N-linked glycosylation sites on bioactivity.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: CHO, Chinese hamster ovary; CTP, carboxy-terminal peptide; hCG, human chorionic gonadotropin; MALDI, matrix-assisted laser desorption ionization; pI, isoelectric point; rhFSH, recombinant human FSH; t1/2, terminal half-life.
Received March 2, 2004.
Accepted July 19, 2004.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-hydroxylase/C17–C20 lyase in the ovaries of immature hypophysectomized rats treated with recombinant FSH combined with various doses of human chorionic gonadotropin. Eur J Endocrinol 137:530–536[Abstract]
This article has been cited by other articles:
 |
|
 |
|
  B.C.J.M. Fauser, B.M.J.L. Mannaerts, P. Devroey, A. Leader, I. Boime, and D.T. Baird Advances in recombinant DNA technology: corifollitropin alfa, a hybrid molecule with sustained follicle-stimulating activity and reduced injection frequency Hum. Reprod. Update, May 1, 2009; 15(3): 309 - 321. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Stork, K. A. Zettlitz, D. Muller, M. Rether, F.-G. Hanisch, and R. E. Kontermann N-Glycosylation as Novel Strategy to Improve Pharmacokinetic Properties of Bispecific Single-chain Diabodies J. Biol. Chem., March 21, 2008; 283(12): 7804 - 7812. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. K. Trousdale, S. V. Pollak, J. Klein, L. Lobel, Y. Funahashi, N. Feirt, and J. W. Lustbader Single-Chain Bifunctional Vascular Endothelial Growth Factor (VEGF)-Follicle-Stimulating Hormone (FSH)-C-Terminal Peptide (CTP) Is Superior to the Combination Therapy of Recombinant VEGF plus FSH-CTP in Stimulating Angiogenesis during Ovarian Folliculogenesis Endocrinology, March 1, 2007; 148(3): 1296 - 1305. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
